The proposed studies will investigate peripheral insulin resistance (high plasma insulin and reduced insulin effectiveness) as an antecedent or risk factor for Alzheimer's disease (AD). In work accomplished during our current funding period, we identified several potential mechanisms through which abnormal glucose and insulin metabolism may contribute to AD pathology and symptoms. We showed that raising plasma insulin levels increased plasma and CSF Abeta42 levels, as it increased CSF levels of pro-inflammatory cytokines and F2-Isoprostanes (markers of brain lipid peroxidation). Importantly, insulin-induced changes in plasma Abeta42 levels were closely related to changes in the inflammatory cytokine, tumor necrosis factor-alpha (TNFalpha). This pattern was most evident in people with higher body mass indices, a recently identified risk for AD. These findings suggest that insulin-induced increases in TNFalpha and free fatty acid (FFA) levels that occur with obesity may affect AS regulation and thereby increase the risk of AD. In contrast, we showed that in mild AD, treatment with me insulin-sensitizing, anti-inflammatory agent rosiglitazone preserves cognition and modulates plasma Abeta levels. A key hypothesis that has emerged from this previous work posits that insulin resistance induces inflammation and FFA elevation, which interact with insulin-associated increases in pathogenetic beta-amyloid (Abeta42). To test this hypothesis, we will conduct three studies examining the effects of induced or improved insulin resistance on cognition, on plasma and CSF Abeta levels, and on inflammatory reactants: 1. ingestion of a high saturated fat/high glycemic index diet for 4 weeks, a time sufficient to induce metabolic changes of insulin resistance; 2. infusion of a triglyceride emulsion, an established procedure for induction of insulin resistance through FFA elevation; and 3. administration of rosiglitazone, a peroxisome proliferator-activated receptor gamma (PPARgamma) agonist which improves sensitivity and reduces inflammation. We predict that induced insulin resistance will increase plasma Abeta42 and CSF inflammatory markers and impair cognition. We further predict that the magnitude of change will be greater for adults with AD, and for those adults with greater central adiposity. We also hypothesize that cognitive enhancement following treatment of insulin resistance with the rosiglitazone will be associated with reduced levels of insulin, FFA, and inflammation. The proposed studies will provide important information about the mechanisms through which peripheral insulin resistance increases the risk of AD, and about therapeutic approaches to ameliorating this risk.